Regex 1

([0-9])\1{2}(?<=\1)(?=\1)

Regex 2

([0-9])\1 (?<=\1)(?=\1)

My input is "43335".

I had an idea why I needed this, but as I was playing around with it, I noticed something strange.

I have a pretty goof understanding why the Regex 2 matches, but I absolutely can not understand why the Regex1 fails.

The difference in behaviour begins when the regex enters the negative lookahead (?=\1)

1. In case of Regex 2 it sees that 5 is not equal to 3 and thus the position is stepped back(1 step) and the first lookaround is entered again (?<=\1)
2. In case of Regex 1 it sees that 5 is not equal to 3 and the position is stepped all the way back to the beginning of the string for some reason.

Could you please help me understand why this happens?

CodePudding user response：

Regex #1:

1. Look at the first character, is it a digit? Yes, it's a 4. Put the 4 into capture group 1. Now look at the next character. Is it a 4. No, it's a 3. No match. Move on to character #2.
2. Look at the 2nd character, is it a digit? Yes, it's a 3. Put 3 into capture group 1. Are the next 2 characters also a 3? Yes, keep going. Now start the lookahead. Is it a 3 also? Ooh, no, it's a 5. This one fails too.
3. Look at the 3rd character. Is it a digit? Yes, it's a 3. Put the 3 into capture group 1. Are the the next two digits also a 3. No, the first one is, but the second one is a 5. No match.
4. ... similar no match for the rest of them.

Regex #2

1. [Same as regex1 step 1] Look at the first character, is it a digit? Yes, it's a 4. Put the 4 into capture group 1. Now look at the next character. Is it a 4. No, it's a 3. No match. Move on to character #2.
2. Put 2nd character (3) into capture group 1. The next 2 characters are the same, try to consume them all, then start the lookahead - is it a 3. No, it's a 5, no match. But now we can backtrack to the \1 and consume only one 3 instead of 2. Move forward again to the first lookahead. Is it a 3? Yes. Back up and do the second lookahead. Is it also a 3? Yes, and we're done, SUCCESS.

What gets matched are the first two 3s.

#1 does not work because you require 4 of the same digits in a row (with the lookahead).

#2 works because you give it the option of 3 consecutive digits as a match.